Treating cancers by targeting their blood vessels is a strategy that has generated great enthusiasm based on its sound scientific rationale and because it has been shown to be highly effective in preclinical tumor studies. However, the results of human clinical trials of this therapy have been less promising. A potential explanation for this disparity is that the vasculature in most human cancers is much less susceptible to antivascular therapy than those in the transplanted mouse tumor models used in preclinical studies. Support for this comes from study of pericytes, which are mesenchymal cells that cover microvessels as part of their maturation process. Pericyte coverage protects and marks vessels that are resistant to antivascular therapy, and the fraction of pericyte-covered vessels is significantly higher in many common human cancers than in transplanted mouse tumors. However, certain autochthonous mouse mammary carcinomas, like those arising in MMTV-neu transgenic mice and mice infected with MMTV, resemble human cancers in having vasculature with extensive pericyte coverage. Importantly, these may also be less susceptible to antivascular agents. Based on this, carcinomas arising in MMTV-neu and MMTV-int-1 transgenic mice are hypothesized to more faithfully model the therapeutic response of common human cancers to antivascular agents than transplanted mouse tumors. Studies proposed in Aim 1 will validate MMTV-neu and MMTV-int-1 tumors as models of human tumor vasculature by comparing vessels and angiogenic activity in mouse and human tumors using a variety of histopathologic techniques. Studies in Aim 2 consist of preclinical trials in MMTVneu and MMTV-int-1 transgenic mice using antivascular agents to treat tumors that have formed, to prevent tumors from forming and in combination with other therapies. Studies in Aim 3 will test manipulation of Tie 2 activity to alter tumor vessel pericyte coverage and response to antivascular therapy using both gene therapy and transgenic approaches to reduce pericyte coverage of MMTV-neu and MMTV-int-1 tumor vessels. Together, these studies will demonstrate whether mammary tumors in MMTV-neu and MMTV-int-1 transgenic mice are superior models of human tumor vasculature which should be used in preclinical evaluation of antivascular agents and strategies to make it more informative and predictive of outcome in patients. Use of such models will enhance development of antivascular agents and make their transition into the clinic more efficient and effective.